Using SentiWordNet and Sentiment Analysis for Detecting Radical Content on Web Forums

This 2012 paper presents a model that was built using sentiment analysis scores from analyzing selected web forums that included radical content. The study looked at two Arabic language web forums and extracted a sample of 500 sentences from each, which were then manually translated. SentiWordNet, Wordnet, and Lexicon were then used to assign positive or negative sentiment to sentences in the sample.

This article will be of use to researchers and practitioners interested in both detecting radical content online and sentiment analysis in general. The internet has become a major tool for communication, training, fundraising, media operations, and recruitment, and these processes often use web forums. The study model utilized by the authors was able to identify more radical content on one forum than the other.